Collapsing 3d printer

[Ryan Carlyle]

I have an idea for a folding (sort of) 3d printer. But not just a Mendel-type architecture that pivots between the XZ stage and Y stage, I mean it extends from a compact shape as it prints. So I think a better word is "collapsing" rather than "folding." 

Aside from SCARA / serial arm type printers, has anybody seen a printer that starts flat and then extends the gantry upwards as it builds Z height? I think people have proposed scissor lift mechanisms for Z stages before?

[Ryan Carlyle]

The big issue with most high-ratio extension mechanisms like scissor lifts, aside from cumulative slop/backlash at rotary joints, is that there's a non-linear relationship between actuator motion and stage motion. So you need to get the kinematics into firmware, and resolution isn't constant. So some cleverness is required...

[whosawhatsis]

I believe some of the early Type A Machines prototypes (shown at Maker Faire before they had anything to sell) used a scissor lift Z mechanism. Unsurprisingly, they apparently never got it to work well enough to make it a product. Unlike most 3D printer makers these days, they didn't try to sell it anyway, which is part of the reason they're still around today.

There's also this one from Japica (one of the biggest jokes in the 3DP community): https://www.kickstarter.com/projects/japica/japicain-revolution-changing-the-way-you-print-in

There are also a couple of telescoping screw mechanisms that have been thrown out as possibilities. Most seem to be bulky enough that you could also use them as linear guides, though they'd need super-high precision to avoid wobble (which would probably be much more stochastic than typical Z wobble because of the variability of the multi-stage screw system).

On Sunday, September 6, 2015 at 09:28, Ryan Carlyle wrote:

The big issue with most high-ratio extension mechanisms like scissor lifts, aside from cumulative slop/backlash at rotary joints, is that there's a non-linear relationship between actuator motion and stage motion. So you need to get the kinematics into firmware, and resolution isn't constant. So some cleverness is required...

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.
To post to this group, send email to 3dp-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/8ee32d17-f008-4092-a056-faea1e9dcc4f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Wing Wong]

Interesting... Was watching one of the how they do it shows about skyscraper constructions and there was this climbing mold for one of the building constructions. Would be an interesting approach for arbitrary height printers that effectively use the printed item as the vertical extension as well... Perhaps something like that, but reusuable? Like a column extrusion system?

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/4BBC574ACE134383AE24461C905614A4%40gmail.com.

[whosawhatsis]

Yeah, I've thought about building a printer that would print a rack for a rack/pinion Z axis as it goes. It was mostly an idea for a special-purpose printer, but it could be something like a rat that is added to all prints. It's hard to think of a way to make it reusable, though.

[Ryan Carlyle]

I love the "print your own Z stage" approach. Awesome concept. Lots of engineering challenges with that though.

WW, you ever thing about taking that super telescoping jig (portable scanner?) and building a Z stage like that? I'm guessing the precision isn't high enough, but I'm sure it crossed your mind?

[whosawhatsis]

I never spent much time thinking about it. In order to maintain stability, you need a fair amount of overlap between stages. If you want it to get very small when not extended, you need a lot of stages, which means the mechanism is going big in at least one dimension. You trade length in the Z direction for size in another dimension. Consider the fact that you'll need at least two and preferably three or four of them, and that's a lot of mass and volume. It might be useful for making a really large-format printer that you can still get into the back of a minivan, but it's probably not a practical way to make a small printer that fits in an even smaller space. For a smallish printer like the Bukito, the mechanical components actually take up a significant amount of the machines's volume. We disconnect the Z axis to make it a bit more flat for travel, but once we add some bubble wrap to keep the parts from smacking into each other, the volume's not significantly less than the assembled machine. Since the platform and X axis hang out over empty space, I actually thought about designing those so that they could be easily removed and stashed within the space in front of the Z beam, which I think would make it collapse more efficiently, but I never found a satisfactory way to do that.

The T-bot printer design I've been working on is almost ideal for folding. With a detachable platform like the one on the Makerbot Mini (5th gen), the X/Y stage hinged/detachable (requires a high-precision quick locking mechanism I haven't quite worked out) and a low-precision folding base just to keep it from falling over while printing, it could easily fold flat.

Oh, that reminds me, the Portabee Go is another printer designed to fold up, and there have also been a couple of brief case printer designs that might have ideas to look at, though all of these do need to be deployed before printing rather than actually expanding as they print.

What you really want is a printer without a platform, though. If you had a printer that could create a layer and then push it through a screen or membrane, then repeat to build up an object, it would essentially be variable-profile extrusion, and the Z axis would be unlimited. I'm imagining something like this: http://makezine.com/2010/09/24/soap-bubble-printer/, but with a digital extrusion die (and printing in a material more useful that soap bubbles). There's not an obvious way to make this work with any existing printing technology, and you'd probably need to invent something that looks like an LCD, but that selectively allows physical material instead of just light to pass through. Some kind of MEMS valve, maybe, or make your entire platform out of piezo inkjet heads. Anyway, you're probably not going to be able to build one capable of any interesting resolution with hobbyist-level equipment any time soon.

I also came up with an idea some time ago for continuous printing on a conveyor belt (http://reprap.org/pipermail/reprap-dev/2011-February/002909.html) that could push one end of a print out of a machine while still adding material to the other end. A month or two later, another thingiverse user (apparently independently) posted a concept design doing the same thing, and about a year later I saw a story that some big industrial 3DP company had apparently patented the process.

As you might have noticed, I've spent a lot of time thinking about this problem from various angles.

On Sunday, September 6, 2015 at 17:30, Ryan Carlyle wrote:

I love the "print your own Z stage" approach. Awesome concept. Lots of engineering challenges with that though.

WW, you ever thing about taking that super telescoping jig (portable scanner?) and building a Z stage like that? I'm guessing the precision isn't high enough, but I'm sure it crossed your mind?

--

You received this message because you are subscribed to the Google Groups "3DP Ideas" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/be21462a-340b-4268-9a70-87d4c3a93afa%40googlegroups.com.

[Ryan Carlyle]

The infinite conveyor w/ slanted slice axis is also a good idea. And for a change, I actually don't think there'd be any patent issue -- the Makerbot ABP patent(s) specifically covers moving the object out of the build volume after completion, or moving multiple print surfaces sequentially through the build volume. Continuously printing one part on a conveyor belt motion stage shouldn't infringe.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[Ryan Carlyle]

Anyway, I was up until 3am designing an extending Z stage mechanism in my head last night. (Couldn't sleep.) Imagine a Foldarap / Mendel style printer, but the Z mechanism progressively tilts up the XZ tower as you print. Yeah, you could do this with a couple super high-torque, high-precision rotary actuators at the pivot point and some complex custom firmware, but I don't roll that way. The three clever bits:

• The X bridge pivots to keep the extruder vertical as the XZ tower rotates up
  
• The Z mechanism lifts the X bridge straight up as the XZ tower rotates up
  
• The Z actuator linkage is linearized, so you don't need special firmware and the Z resolution is constant
  

It'll run standard cartesian firmware and should be "plug and print." I'm not expecting amazing Z height precision or flawless XY position registration, but I think it'll work well enough to print stuff. The big design challenge is actually more around making high-precision revolute joints and stiff linkages than the drivetrain concept itself. 

I need to wrap up some design on the Tesseract Delta rev2 drivetrain, and then I'll put together concept drawings. (Although I bet WW will invent a similar-but-different way of doing it before I get renders posted.)

[ekaggrat singh kalsi]

https://plus.google.com/+NicholasSeward/posts/d2vhGGjsFVt

[Ryan Carlyle]

Good one, thanks for posting that. I'm also using Scott-Russel linkages, but I think I've got a more elegant solution. Nick loves making printers that require custom firmware, are hard to tram, and use as many wobbly joints as possible :-)    

(I'm a big fan of his work, we're just working with different design goals in mind)

[Ryan Carlyle]

On Sunday, September 6, 2015 at 9:42:43 PM UTC-5, Ryan Carlyle wrote:

use as many wobbly joints as possible 

On thinking about it, this isn't quite fair, since I'll have a couple more joints in my design than Nick used in that one. 

[ekaggrat singh kalsi]

i had designed a folding printer a year back and prototyped it . but was not happy with it aesthetically and also getting it aligned was a little problem.. laser cut plywood is not the best of the materials ...

http://www.thingiverse.com/thing:298544

[whosawhatsis]

On Sunday, September 6, 2015 at 19:25, Ryan Carlyle wrote:

• The X bridge pivots to keep the extruder vertical as the XZ tower rotates up
  
• The Z mechanism lifts the X bridge straight up as the XZ tower rotates up
  
• The Z actuator linkage is linearized, so you don't need special firmware and the Z resolution is constant
  

Well, keeping the X bridge parallel to the base is easy, you just need a four-bar (parallelogram) linkage.

Making the X bridge lift straight up (rather than in an arc) is as simple as making the long arm of the scott-russell linkage exactly twice as long as the short one and fixing the point at the bottom of the short one so that the long end always stays above it.

Off the top of my head, the simplest way to linearize the actuator would be to use an extension spring to pull the two linkage ends on the base together horizontally, and spool out a cord or belt vertically from a motor (or around an idler) mounted directly under the top beam. Let out N mm of cord and spring pulls the linkage ends together enough allow the the X bridge to go up N mm. You'd definitely want a geared motor running the Z axis to deal with the spring force, and also to give the unpowered motor enough resistance to keep the thing from popping up when you cut power (in fact, you'd probably want to disconnect the spring for storage/transport, in addition to latching the mechanism in the down position).

I'm sure there's a better way, but that's the only one that comes to me at the moment. I might be up until 3am working on the problem in my head tonight too.

[Ryan Carlyle]

Roughing up the basic concept... Looks a little more complex on paper than it did in my head :-)

Basically, each Scott-Russel linkage in the Z lift is doubled up in a pantograph arrangement. The motion of one pantograph leg is mirrored in its other leg, except rotated 90 degrees from vertical to horizontal. Input = output. So the Z actuator is horizontal, and has linear resolution, and (barring major backlash/friction through the ass-awkward mechanical advantages) has linear lifting force at the X-bridge. 

Yes, the links shown here are L-shaped, I did take that into account in the arrangement of joints. Makes it collapse cleaner. The range of linkage angle shown here is 8.1 to 53.1 degrees (45 degrees net travel) But there's a strong argument to be made for traveling from 22.5 to 67.5 degrees to minimize the mechanical advantages/disadvantages at the extreme ranges of the linkage.

I'd probably build it with a mix of 1515 and lasercut plates instead of the 2020 shown here. That would decrease the thickness a lot. What I have drawn here is quite bulky but could plausibly get you 12x12x6" collapsed size (using a low-profile hot end) with something resembling 6x6x6" build volume. 

The X and Y drivetrains would just be typical Mendel/Prusa style linear stages. Although you could get clever with the X bridge and mount the X stepper on one of the sliding pantogram carriages, with an L-shaped loop belt to move the extruder carriage. That would significantly decrease the mass that has to be lifted through the pantogram. I'm also imagining putting a counterweight on the "back side" X belt to move opposite the extruder and reduce side-to-side shaking. There's also various ways the X axis rigidity of the lift linkages can be increased, but that's just detail work.

[whosawhatsis]

Ah, I hadn't thought of that. Wow, that's a lot of linear rails...

Speaking of a pantograph, you could make the horizontal arms half the length of the vertical ones so that they would only travel half way across. This would allow you to bring both sets of horizontal linkages into the same plane, cutting the height of the mechanism and also letting to get rid of one of the linear rails (instead using two carriages on the same rail). You'd halve the resolution of the Z axis, but I'd want to use a geared motor there anyway (I've never liked the resolution of belt-driven Z). OTOH, you could make that a screw-driven axis, but you'd lose the easy ability to drive both directions at once unless you had a custom screw made that was half left-handed (actually, with a large enough lead/diameter ratio, I bet it would be possible to make a dual thread screw that would work with both left- and right-handed nuts along its entire length... that would be crazy!)

Don't forget the parallelograms, or did you have a different plan for handling that?

I don't think you need to worry too much about a low-profile hot end. Just design it so that the platform has to move forward for the Z to fold all the way down, and the hot end can dip below it for transport. Even if you switch everything (including the bottom frame) to 1515, you'll have more than 60mm of height fully folded once the parallelograms are added, and an e3d v6 is 62.3mm long. If you have a bowden coupler at the very top of that that you disconnect for transport, you shouldn't have any trouble stashing the hot end inside when it folds up. If you keep the 2040 for the bottom frame, that should give you more than enough space for a direct-drive extruder with a MK7-style hot end if you arrange the X axis carefully.

--

You received this message because you are subscribed to the Google Groups "3DP Ideas" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/935fd3f8-d2b4-404e-a824-9dc03d1ac70b%40googlegroups.com.

[ekaggrat singh kalsi]

http://www.3ders.org/articles/20150906-foldie-3d-fff-printer-can-be-easily-carried-around-in-just-a-19-inch-laptop-bag.html

[whosawhatsis]

You know, since you're putting the X bridge all the way at one end of the machine anyway, you could forgo the moving bed and use one of my cantilevered T-shaped XY stage designs instead...

On Sunday, September 6, 2015 at 23:09, Ryan Carlyle wrote:

--

You received this message because you are subscribed to the Google Groups "3DP Ideas" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/935fd3f8-d2b4-404e-a824-9dc03d1ac70b%40googlegroups.com.

[whosawhatsis]

Oh, and if you go with a screw drive, you could still use a belt with idler pulleys on both ends (or a simple cable loop) to keep the two reverse carriage pair in sync.

1:30 now. Looks like the 3am thing is likely. I'm building some crazy stuff in my head.

[whosawhatsis]

Simplify it. Use only one of the linearizing pantograph mechanisms and use a passive cable drive to keep the two horizontal carriages in sync.

[Ryan Carlyle]

On Monday, September 7, 2015 at 2:45:29 AM UTC-5, Whosa whatsis wrote:

Ah, I hadn't thought of that. Wow, that's a lot of linear rails...

It is, but they're short :-) just have to buy a long cheap one and cut it up. It IS a lot of carriages. 

 

Speaking of a pantograph, you could make the horizontal arms half the length of the vertical ones so that they would only travel half way across. This would allow you to bring both sets of horizontal linkages into the same plane, cutting the height of the mechanism and also letting to get rid of one of the linear rails (instead using two carriages on the same rail). You'd halve the resolution of the Z axis, but I'd want to use a geared motor there anyway (I've never liked the resolution of belt-driven Z).

I think the mechanical advantages are already going to cause a lot of flex at the extreme positions, adding additional ratios doesn't seem like a great idea. Also, without making the linkage more complex, I'm not entirely sure a 2:1 reduction would stay linear from actuator to output. Maybe it would? Not sure. 

 

OTOH, you could make that a screw-driven axis, but you'd lose the easy ability to drive both directions at once unless you had a custom screw made that was half left-handed (actually, with a large enough lead/diameter ratio, I bet it would be possible to make a dual thread screw that would work with both left- and right-handed nuts along its entire length... that would be crazy!)

You ever seen a self-reversing screw? We use them for levelwinds on large umbilical/wire winches. GIS link

The normal operation is that a rotating pawl engages the screw thread, and the pawl turns to match the opposite pitch when it hits the end of travel. But I guess you could set up two carriages on the same screw, one on either side, and get both directions of motion.  

I think avoiding exposed grease/oil needs to be a pretty significant design objective for something that travels in a backpack -- so I'm hesitant to use screws. Belts, delrin v-wheels, and igus rails seem like the best way to go. We can use geared steppers to get the resolution. And maybe wider/stronger belts for better precision.

Don't forget the parallelograms, or did you have a different plan for handling that?

I haven't decided between a delta-style parallelogram or a passive belt + large pulley arrangement. The belt gives you some advantage for tensioning slop out of the revolute joints, and probably ends up being lower parts count than the parallelogram linkage.

 

I don't think you need to worry too much about a low-profile hot end. Just design it so that the platform has to move forward for the Z to fold all the way down, and the hot end can dip below it for transport. Even if you switch everything (including the bottom frame) to 1515, you'll have more than 60mm of height fully folded once the parallelograms are added, and an e3d v6 is 62.3mm long. If you have a bowden coupler at the very top of that that you disconnect for transport, you shouldn't have any trouble stashing the hot end inside when it folds up. If you keep the 2040 for the bottom frame, that should give you more than enough space for a direct-drive extruder with a MK7-style hot end if you arrange the X axis carefully.

Yes, this is all true. I think I'd like to capture the usable Z travel though. If the hot end stows below the bed, that's taking away from your printable range of motion. E3Dv6 is quite long compared to what I can rig up with a Rep2 style cooling bar. Just need to run the heatsink horizontal instead of vertical. I'm also wondering if I could get away with putting a "gooseneck" guide arch into the bowden tube so it comes off the extruder carriage horizontal and doesn't need to be detached. (Will cause PLA to break in storage, of course.)

 

On Monday, September 7, 2015 at 3:41:32 AM UTC-5, Whosa whatsis wrote:

Simplify it. Use only one of the linearizing pantograph mechanisms and use a passive cable drive to keep the two horizontal carriages in sync.

 This seems like a good simplification, but I worry about adding a lot of slop/flex through the sync mechanism. The biggest challenge with this design is keeping the backlash and play down. Might be a better approach though... the crossing linkages add a lot of complexity. 

[Ryan Carlyle]

On Monday, September 7, 2015 at 2:47:52 AM UTC-5, ekaggrat singh kalsi wrote:

http://www.3ders.org/articles/20150906-foldie-3d-fff-printer-can-be-easily-carried-around-in-just-a-19-inch-laptop-bag.html

WW and I were discussing that on G+ yesterday. It has a lot of assembly required to get it running. Nice approach though. 

[Ryan Carlyle]

Here's the simplified, low-profile version with some of WW's suggestions:

I'm not at all committed to the MGN12 and belt drive for the Z actuator. A belt + V-wheels would avoid grease mess, but a linear rail and screw would probably be more precise. (I'm not a huge fan of V wheels on 2020, and it's arguable whether adding envelope for 2040 is worthwhile here.) Lots of options. Just needs very high bending stiffness (no 8mm round rods). 

[whosawhatsis]

I wasn't thinking about using a cross-beam. It looks like the tension on the cable would make the beam want to skew. If you designed it to be skewed the other way a few degrees, the tension would push the wheels against the frame sides, and you may even be able to make it work without the outer wheels. You could also just balance the forces by adding a second cable as a mirror image of the first. Because the relative Y-axis positions of the two pulleys is constant and the pulley rotation direction is the same, you could do this by running both cables around the same pulleys (crossing around each pulley, and again in the middle) so that you don't need more pulleys to add the second (mirror image) cable.

What I was thinking of was a cable in a figure-8 loop, so that the two sides would move together and the forces would be balanced. In that case, the cross beam shouldn't be necessary (though it would make the initial syncing of the two sides easier, and aside from adding inertia, it shouldn't hurt anything). This would require four idler pulleys, while as I mentioned, you could do your version plus its mirror image with only two.

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/4e92b9a0-3acd-4731-aa16-5697688a9628%40googlegroups.com.

[Ryan Carlyle]

I actually added the cross-beam because it provides a huge amount of rotational stiffness to the carriages, which makes it much easier to keep the X-bridge from swaying side to side. Rotation around the axis of the extrusion is the biggest issue I've had with V-wheels. Good points on the other stuff though.

On Monday, September 7, 2015 at 5:04:50 PM UTC-5, Whosa whatsis wrote:

I wasn't thinking about using a cross-beam. It looks like the tension on the cable would make the beam want to skew. If you designed it to be skewed the other way a few degrees, the tension would push the wheels against the frame sides, and you may even be able to make it work without the outer wheels. You could also just balance the forces by adding a second cable as a mirror image of the first. Because the relative Y-axis positions of the two pulleys is constant and the pulley rotation direction is the same, you could do this by running both cables around the same pulleys (crossing around each pulley, and again in the middle) so that you don't need more pulleys to add the second (mirror image) cable.

What I was thinking of was a cable in a figure-8 loop, so that the two sides would move together and the forces would be balanced. In that case, the cross beam shouldn't be necessary (though it would make the initial syncing of the two sides easier, and aside from adding inertia, it shouldn't hurt anything). This would require four idler pulleys, while as I mentioned, you could do your version plus its mirror image with only two.

On Monday, September 7, 2015 at 14:33, Ryan Carlyle wrote:

Here's the simplified, low-profile version with some of WW's suggestions:

I'm not at all committed to the MGN12 and belt drive for the Z actuator. A belt + V-wheels would avoid grease mess, but a linear rail and screw would probably be more precise. (I'm not a huge fan of V wheels on 2020, and it's arguable whether adding envelope for 2040 is worthwhile here.) Lots of options. Just needs very high bending stiffness (no 8mm round rods). 

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[Ryan Carlyle]

It occurred to me today that putting a cross-beam between the two tilt-arms would probably be a much better approach than between the two carriages. Turning the long arms on the two lift linkages into a rigid metal square would add a ton of stiffness against both X-wobble and uneven lift height. The downside is machine envelope -- I was hoping to put the build plate at the same height as that cross-beam would be. I think there's workarounds there but it would take some careful layout.

[JasonB]

I ran in to this last week. 

https://www.youtube.com/watch?v=xT3xcY1s1kQ

https://www.youtube.com/watch?v=801ZII2e3CM

It's ridiculously clever, and who knows if it can be scaled down to a reasonable size for a printer and still maintain that high side loading strength.  It would also be outrageously expensive.  But it would look amazing.

[Ryan Carlyle]

Wow. That's freakin' awesome. It's practically assembling a lead screw from parts each time it extends.

I wonder how many cycles it can do before the metal starts to fatigue.

[Carl]

I wonder how many cycles it can do before the metal starts to fatigue.

Around 25000 cycles...

[Ryan Carlyle]

Another idea that comes to mind for a collapsing printer is some flavor of grounded delta. IE rotary-actuator delta with the arm pivots below the build plate. With detachable joints like mag-balls, this could pack away to be extremely compact.

[Ryan Carlyle]

Now that my tesseract delta is done, I need to decide if I want to work on upgrades/maintenance of my existing printer "fleet" or start real design work on this collapsing printer idea. I think I want to build it, I'm just not necessarily in the market for another printer right now... 

[Ryan Carlyle]

I'm working on designing this now. Can't get the idea out of my head until I build it.

Current plan:
- 4-5" cube working volume
- 300x250mm footprint to fit in a backpack (a little bigger than a textbook)
- Trying for <100mm thickness collapsed, doubt I'll pull it off though. 125-150mm would probably be fine.
- Structural frame a mix of 2020 series extrusions (eg probably some 20x80 in there)
- All moving linkages will be 1515, connected with printed parts and probably 623 bearings
- One side of the Z lift will have a full pantograph with the other side most likely cable-sync'd per WW's suggestions
- MGN9 rails for Z, Misumi PLRH for Y bed, not sure about X yet -- thinking mini V wheels or another MGN9 or even a couple 8mm rods just because I have them on hand
- Short-throw Bowden, probably a B'Struder
- Unheated build plate

I'm not too far into it yet, mostly just working on general mechanism layout right now. Fitting all the parts into this kind of envelope is hard.

[whosawhatsis]

I haven't tried them, but DeltaPrintr sells wheels designed to roll on 1515 extrusions: http://deltaprintr.com/shop/parts-and-accessories/hardware/carriage-wheel/. They're smaller than even the mini-V wheels and would save you from needing to use a different type of extrusion just for one piece.

--

You received this message because you are subscribed to the Google Groups "3DP Ideas" group.

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/db67426d-6c42-44b0-9998-7fac91bc1092%40googlegroups.com.

[Tempo502]

Oh yeah, forgot about those, they're pretty perfect. 

[Ryan Carlyle]

Here's another neat "collapsing linear actuator" concept I just found. Basically you push anti-backbend chain out of a storage cartridge. Combined with a linear rail or another chain oriented the opposite direction, it gets you a linear stage. I doubt it's precise enough for standard 3d printing, but it's a possible candidate for the "self-printing Z-stage" type actuator. The printer could literally construct more chain links as it extends. 

https://en.wikipedia.org/wiki/Rigid_chain_actuator

[JasonB]

Oh, that's even better.  Following some references brought me to this implementation. 

It's timing belts and plastic, pretty much perfect for 3d printing.  Combine with a clever use of the aforementioned 1515 wheels for linear constraint and you might have a z stage.

On Saturday, April 16, 2016 at 2:11:31 PM UTC-7, Ryan Carlyle wrote:

Here's another neat "collapsing linear actuator" concept I just found. Basically you push anti-backbend chain out of a storage cartridge. Combined with a linear rail or another chain oriented the opposite direction, it gets you a linear stage. I doubt it's precise enough for standard 3d printing, but it's a possible candidate for the "self-printing Z-stage" type actuator. The printer could literally construct more chain links as it extends. 

https://en.wikipedia.org/wiki/Rigid_chain_actuator

[Ryan Carlyle]

Benefit to chains is that you can print them with the same rigid plastic as your main print. A belt like this would need to be printed with a flexible filament. (Which is fine for a regular collapsing printer but not a self-extending printer). Would probably be higher resolution and more compact when retracted though. What I'm not sure about is how rigid/precise the belt version is under load. 

[JasonB]

Stroke:  from  100  mm to many  meters  with  or  without
guide
Max. static load : 500 kg*
Elongation**: 3 mm / 100 kg
Compression**: 1 mm / 100 kg
Weight of a belt: 700g/m
Section of a belt: 25 mm x 28 mm
Note: 
*Maximum value without safety coefficient for R&D application.
**at  max.  load.  This information  is  based  on  static  testing  performed  by an independent laboratory. Dynamic capacity and life cycle depend on the drive design.
Capacity: 10 N to 1000 N
Speed: 20 to 300 mm/s

1mm compression at 100kg doesn't seem to be much to worry about.  Except if you're going to print one for this printer it will be significantly smaller so maybe some built in preload in the belts is best; in other words, a slight interference fit between teeth. 

True, you probably won't be printing this whole belt as you go but you could have reserve belt in the design and print teeth as needed perhaps.  That might require removable teeth interlocked to the belt.  This gets too complicated I think.  But damn I love the idea of watching that go.

[Ryan Carlyle]

I'm guessing they're using some extremely high-durometer rubber and a pretty beefy set of belts, to get that kind of elasticity performance. 

[Ryan Carlyle]

Relevant: http://spectrum.ieee.org/automaton/robotics/robotics-hardware/spiral-zipper-joint-creates-robot-arm-out-of-a-strip-of-plastic